SO2 conversion to sulfones: Development and mechanistic insights of a sulfonylative Hiyama cross-coupling

Article abstract of DOI:10.1039/c9cc06858a

A Pd-catalyzed Hiyama cross-coupling reaction using SO₂ is described. The use of silicon-based nucleophiles leads to the formation of allyl sulfones under mild conditions with a broad functional group tolerance. Control experiments coupled with DFT calculations shed light on the key steps of the reaction mechanism, revealing the crucial role of a transient sulfinate anion.

Full text of DOI:10.1039/c9cc06858a

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
SO2 conversion to sulfones: development and
mechanistic insights of a sulfonylative Hiyama
cross-coupling†
Cite this: Chem. Commun., 2019,
55, 12924
Received 3rd September 2019,
Accepted 3rd October 2019
Aurelien Adenot, Joelle Char, Niklas von Wolff,
Guillaume Lefevre,
`
´
¨
Lucile Anthore-Dalion
and Thibault Cantat
*
DOI: 10.1039/c9cc06858a
rsc.li/chemcomm
A Pd-catalyzed Hiyama cross-coupling reaction using SO2 is described.
Shortly after the discovery of the eponymous coupling,
The use of silicon-based nucleophiles leads to the formation of Hiyama et al. reported the carbonylative coupling of aryl iodides
allyl sulfones under mild conditions with a broad functional group with organosilanes in the presence of a palladium catalyst (Fig. 1a).11
tolerance. Control experiments coupled with DFT calculations shed SO2 is both more electrophilic and nucleophilic than CO,12 and its
light on the key steps of the reaction mechanism, revealing the frontier orbitals are centered on the sulfur atom as they are on the
crucial role of a transient sulfinate anion.
carbon atom of carbon monoxide (Fig. 1c). Besides, the migratory
insertion of SO2 in a Pd–C bond has already been reported by
Present in many contemporary pharmaceuticals, agrochemicals Goddard and co-workers (Fig. 1b).13 We hence hypothesized the
and materials (e.g. the antibiotic Thiamphenicol or the herbicide feasibility of a sulfonylative Hiyama cross-coupling.
pyroxasulfone), sulfones are also used as key intermediates in
We began our investigation by exploring the coupling of
organic synthesis1,2 (e.g. the Julia olefination3 or the Ramberg– triethoxy(allyl)silane (1a), 4-iodotoluene (2), and gaseous sulfur
4
¨
Backlund reaction ). Given this combination of a prominent dioxide, generated by thermal decomposition of K2S2O5 in a
biological activity and an appealing synthetic utility, numerous two-chamber apparatus (see ESI†). In the presence of Pd(acac)2
methodologies have been developed for their preparation.1 as a catalyst and TBAFꢀ3H2O as a fluoride source to activate the
Because it has a high atom-efficiency, the insertion of a sulfur weakly polar C–Si bond, the desired sulfone 3a was obtained in
dioxide molecule upon coupling a nucleophile with an electrophile 12% yield (Table 1, entry 1). By contrast, the bench stable
has recently emerged as a valuable route.5 Organomagnesium,6 surrogate of SO2, DABSO, popularized by Willis et al.,14 gave
organozinc7 and organoboron8 compounds were successively no desired product (Table 1, entry 2), presumably due to the
reported as nucleophiles; yet, they suffer from toxicity issues, coordination of the DABCO by-product to palladium. Changing
functional group incompatibility, and/or air-sensitivity.9
the fluoride source to the anhydrous tetrabutylammonium
Because they are readily available, air-stable and show an difluorophenyl-silicate (TBAT) improved the reaction efficiency
improved functional-group tolerance, organosilanes were recently (17% yield, Table 1, entry 3). While two equivalents of SO2
considered to produce sulfones from SO2 or SO2 surrogates.10 increased the yield up to 26% (Table 1, entry 4), an excess of SO2
However, up-to-date methods are still limited to sp3-hybridized was detrimental to the reaction (Table 1, entry 5), possibly
electrophiles, which react through S-alkylation after the formation because of a poisoning of the catalyst. After screening a variety
of an intermediate sulfinate anion (Scheme 1). Unlocking the
utilization of sp2-hybridized electrophiles would require a change
of mechanism; and, to tackle this issue, we report herein the
first sulfonylative Hiyama cross-coupling affording sulfones from
organosilanes, sulfur dioxide and aryl iodides, in a single-step
reaction (Scheme 1). Mechanistic control experiments, combined
with DFT calculations performed on the key reaction steps,
provide insight into the mechanism of the reaction.
´
NIMBE, CEA, CNRS, Universite Paris-Saclay, CEA Saclay, 91191 Gif-sur-Yvette,
France. E-mail: thibault.cantat@cea.fr
Scheme 1 Representative state-of-the-art of sulfone synthesis from SO2
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
and organometallic compounds.
c9cc06858a
12924 | Chem. Commun., 2019, 55, 12924--12927
This journal is ©The Royal Society of Chemistry 2019

View Article Online
Communication
ChemComm
Table 2 Formation of the diallyl sulfone (4) as a side-product
Y3
Yield 3a (%) Yield 4 (%) 4/3a ratio DGa (kcal molꢁ1
)
(OMe)3 (1b)
(OEt)3 (1a)
Me2(OMe) (1c) 74
51
78
23
10
7
0.45
0.13
0.09
0
2.98
3.60
11.94
15.60
Fig. 1 (a) First example of a carbonylative Hiyama cross-coupling; (b)
Me3 (1d)
54
0
insertion of SO2 in a Pd–C bond; (c) frontier orbitals of CO and SO2.
a
Gibbs free energy variation for the fluoride transfer from Me3SiF2ꢁ to
allylsilanes 1 (see ESI for computational details). Yields measured by
1H NMR (internal standard: mesitylene).
Table 1 Influence of the reaction conditions on the sulfonylative Hiyama
cross-coupling of triethoxy(allyl)silane (1a) with 4-iodotoluene (2) (see
Table S1 for a more exhaustive table, ESI)a
(3d–3e, 38–45%, Scheme 2a), as reflected by a Hammett corre-
lation with a slope of r = ꢁ0.39 (Scheme 2b). Alkenyl sulfone 3f
was also prepared in 57% yield (Scheme 2a). Using aryl bro-
mides, diallyl sulfone (4) was exclusively formed, presumably
due to the more difficult activation of the C–Br bond by
oxidative addition.
SO2 source
(eq.)
Yield in
3a (%)
Entry Fluoride source (eq.)
Ligand (%)
As regards the nucleophile, methyl-substituted allylsilanes
1e–1g successfully provided the desired sulfones in 33–74%
yields, with the selective formation of the a-substituted sulfones
from the corresponding g-substituted allylsilanes, while the
classical Hiyama cross-coupling reaction using substituted allyl-
silanes usually faces regioselectivity issues.15 Disappointingly,
triethoxy(phenyl)silane and triethoxy(vinyl)silane exhibited no
reactivity (Scheme 3).
These surprising results prompted us to conduct mecha-
nistic control experiments. The catalytic system is an efficient
catalyst in Hiyama couplings and, in the absence of SO2, 1-allyl-
4-methylbenzene (6a) and biphenyl (6b) were formed from
triethoxy(allyl)silane (1a) and triethoxy(phenyl)silane (5), respectively
(Scheme 4a). Nevertheless, in the presence of SO2, sulfone 3a was
obtained in 78% yield, while the corresponding diarylsulfone 7
b
1
2
3
4
5
6
7
8
9
TBAFꢀ3H2O (1)
TBAFꢀ3H2O (1)
TBAT (1)
TBAT (1)
TBAT (1)
TBAT (1)
TBAT (1)
TBAT (1)
SO2 (1)
—
—
—
—
12
0
17
26
0
DABSO (0.5)
b
SO2b (1)
SO2b (2)
SO2b (4)
SO2b (2)
SO2b (2)
SO2b (2)
—
XPhos (10)
41
Xantphos (10) 78
Xantphos (5) 65
Xantphos (10) 71
TBAF (1 M in THF) (1) SO2 (2)
a
b
Results obtained in THF at 80 1C during 4 h on a 0.1 mmol scale.
SO2 was generated by thermal decomposition of K2S2O5, see ESI.
TBAF = tetrabutyl ammonium fluoride; TBAT = tetrabutylammonium
difluorophenyl-silicate. Yields measured by 1H NMR (internal standard:
mesitylene).
of palladium sources and phosphine ligands (Table 1, entries
5–7 and Table S1, ESI†), a set of conditions (Pd(acac)2 5 mol%,
Xantphos 10 mol%, TBAT 1 eq.) gave the best results in our
hands, yielding the desired sulfone 3a in 78% yield (Table 1,
entry 7). No reaction took place without the catalyst, even after
24 h (Table S1, ESI†).
During the screening process, the diallyl sulfone (4) was
identified as a side-product. The relative proportion of 4 was
found to depend directly on the nature of the substituents at the
silicon atom (Table 2): the quantity of diallyl sulfone (4) increases
with the fluoride-affinity of the organosilane (computed by thꢁe
Gibbs free energy variation for the fluoride transfer from Me3SiF2
to the allylsilane). As a result, tri(ethoxy)allyl silanes were selected
to explore this new reaction, as it provides the best balance
between selectivity and productivity (see ESI†).
The chosen reaction conditions enabled the synthesis of allyl
arylsulfones 3a–3e from aryl iodides bearing electron-donating
substituents as well as electron-withdrawing in 38–82% yield.
The reaction tolerates well the presence of a ketone group and 3e
was formed in 45% yield. Interestingly, sulfones bearing electron-
donating substituents 3a–3b were obtained in better yields
(78–82%) than the ones bearing electron-withdrawing substituents
Scheme 2 (a) Substrate scope in organoiodides; (b) Hammett plot of the
reaction. Reaction conditions: electrophile (1 eq.), allylSi(OEt)3 (1a, 1.1 eq.),
electrophile (1 eq.), SO2 (2 eq., generated by thermal decomposition of
K2S2O5, see ESI†), TBAT (1 eq.), Pd(acac)2 (5 mol%), Xantphos (10 mol%),
THF, 80 1C, 4 h. Yields measured by 1H NMR (internal standard: mesitylene).
Yields of isolated products from scaled-up experiments (1 mmol scale) are
given within parentheses.‡
This journal is ©The Royal Society of Chemistry 2019
Chem. Commun., 2019, 55, 12924--12927 | 12925

View Article Online
ChemComm
Communication
To probe the role of transient sulfinate anions, the organo-
silane nucleophiles 1a and 5 were reacted with SO2 and
ethyliodide, in the absence of the palladium catalyst. While
sulfone 9a was formed in 85% yield from the allylsilane, only
4% of the ethylphenylsulfone 9b were observed, suggesting that
the formation of phenylsulfinate from triethoxy(phenyl)silane
(5) is blocked (Scheme 4b). This finding was confirmed when
diarylsulfone 7 was obtained in 24% yield from the preformed
arylsulfinate 8c exposed to phenyl iodide, in the presence of
the palladium catalyst (Scheme 4c). All together, these data
support path B in Scheme 4d, with the metal catalyst enabling
the coupling between a transient sulfinate anion and the
arylhalide electrophile.
Scheme 3 Substrate scope in organosilanes. aIn the case of aryl- and
vinylsilane, the nucleophile were fully recovered. Reaction conditions:
organosilane (1.1 eq.), 4-iodotoluene (2, 1 eq.), SO2 (2 eq., generated by
thermal decomposition of K2S2O5, see ESI†), TBAT (1 eq.), Pd(acac)2
(5 mol%), Xantphos (10 mol%), THF, 80 1C, 4 h. Yields measured by 1H NMR
(internal standard: mesitylene). Yields of isolated products from scaled-up
experiments (1 mmol scale) are given within parentheses.‡
The regioselectivity observed with substituted allylsilanes
(Scheme 3) readily derives from this mechanism. With an allylic
nucleophile, the sulfinate formation can proceed through two
different mechanisms: either a bimolecular electrophilic sub-
stitution at the a position of the silane (TS1, Scheme 5), or the
corresponding SE20 mechanism with formation of the C–S bond
at the g position (TS2, Scheme 5).§ DFT calculations revealed
that the SE20 path (DG‡(TS2) = 25.4 kcal molꢁ1) is favored over
the corresponding ipso reaction (DG‡(TS1) = 29.7 kcal molꢁ1).
Besides in the case of trimethyl(aryl)silane, for which only TS1 can
be considered, the transition state lies 34.0 kcal molꢁ1 higher
than the starting materials, which explains its lack of reactivity.
In classical Hiyama cross-coupling reactions, the rate-
determining step (RDS) is usually the transmetallation.17 Here,
the pre-formation of a sulfinate anion circumvents such a
transmetallation and the Pd-catalyst instead mediates the
formation of the second C–S bond, connecting the electrophile
and the SO2 fragment. According to the Hammett plot (Scheme 2b),
the rate-determining step of the catalytic reaction is facilitated
with electron-rich electrophiles (r o 0), suggesting that the
rate-determining step is the reductive elimination,18 rather
than the oxidative addition.19
was not observed starting from phenylsilane 5. These observa-
tions demonstrate that the successful formation of sulfones
3a–3i cannot be explained by a direct transposition of the
mechanism reported for carbonylative cross-couplings,16 where
the insertion of the small molecule occurs after the oxidative
addition (path A, Scheme 4d).
Careful 1H NMR monitoring of the reaction revealed the
formation of the allylsulfinate anion 8a as an intermediate
(Fig. S3, ESI†). Another pathway was hence devised, where 8a is
involved in a ligand exchange after an oxidative addition step
(path B in Scheme 4d). The desired sulfone 3 would then be
obtained through a reductive elimination step.
In fact, the energy barrier computed for the reductive
elimination of an allylsulfone from (allylSO2)-Pd(PMe3)2-Ph is
high, at 37.1 kcal molꢁ1 (compared to 23.2 kcal molꢁ1 for allyl-
Pd(PMe3)2-Ph). Importantly, the DFT calculations also point to the
positive influence of electron-donating groups on the kinetics
of this step, in agreement with the slope of the experimental
Scheme 4 Mechanistic control experiments: (a) comparison between
Hiyama cross-coupling and the sulfonylative version; (b) in situ sulfinate
formation; (c) direct cross-coupling of aryl sulfinate 7c with phenyl iodide
(DMF was used for solubility of 8c, see ESI† for more details); (d) proposed
pathways for the mechanism of the reaction. Yields measured by H NMR
(internal standard: mesitylene).
Scheme 5 Computed reaction pathways for the formation of the sulfi-
nate via SE2 or SE20. Level of theory: B3LYP/G-D3/6-311+G(d) (C, H, O) and
6-311++G(d,p) (F, S, Si), PCM was used for THF solvation. Values given
correspond to Gibbs free energies with respect to the starting materials
(G = 0.0 kcal molꢁ1). ‘‘Fꢁ’’ stands in fact for the anion FSO2ꢁ.¶
1
12926 | Chem. Commun., 2019, 55, 12924--12927
This journal is ©The Royal Society of Chemistry 2019

View Article Online
Communication
ChemComm
Notes and references
‡ As shown in entry 9 of Table 1, TBAF (1 M in THF) is a competent
fluoride source and it was used for scaled-up experiments to avoid the
formation of the allyl phenyl sulfone 3c from TBAT.
§ As shown in the ESI,† the unimolecular pathway has to be discarded,
the fluoride transfer from the fluoride source to the organosilane and
subsequent C–Si bond scission of the hypervalent species being too
energetically demanding.
¶ As already reported,10d and experimentally evidenced (see ESI†), the
fluoride anion is actually transferred to SO2 to yield the stable anion
ꢁ
FSO2 which acts as the fluoride source.
1 G. Manolikakes, N.-W. Liu and S. Liang, Synthesis, 2016, 1939.
2 B. M. Trost and C. A. Kalnmals, Chem. – Eur. J., 2019, 25, 11193.
3 M. Julia and J.-M. Paris, Tetrahedron Lett., 1973, 14, 4833.
Fig. 2 Representations of the HOMO in the transition states for the
reductive elimination from (allylY)-Pd(PMe3)2-Ph (with Y = SO2 or S, hydro-
gens and methyl groups from PMe3 were omitted for clarity) and Gibbs free
energy of the transition states (DG‡(TS)) with respect to the starting material
(G = 0.0 kcal molꢁ1). Level of theory: B3LYP/G-D3/6-311+G(d) (C, H, O,
N, F), 6-311++G(d,p) (P, S) and SDD (Pd), PCM was used for THF solvation.
¨
4 L. Ramberg and B. Backlund, Ark. Kemi, Mineral. Geol., 1940, 13, 50.
5 D. Zheng and J. Wu, Sulfur Dioxide Insertion Reactions for Organic
Synthesis, Springer, Singapore, 2017, ch. 3.
6 A. S. Deeming, C. J. Russel, A. J. Hennessy and M. C. Willis, Org.
Lett., 2014, 16, 150.
7 B. N. Rocke, K. B. Bahnck, M. Herr, S. Lavergne, V. Mascitti,
C. Perreault, J. Polivkova and A. Shavnya, Org. Lett., 2014, 16, 154.
8 Y. Chen and M. C. Willis, Chem. Sci., 2017, 8, 3249.
9 Y. Nakao and T. Hiyama, Chem. Soc. Rev., 2011, 40, 4893.
10 (a) L. Bouchez and P. Vogel, Synthesis, 2002, 225; (b) D. Zheng,
M. Chen, L. Yao and J. Wu, Org. Chem. Front., 2016, 3, 985;
(c) D. Zheng, R. Mao, Z. Li and J. Wu, Org. Chem. Front., 2016,
3, 359; (d) N. von Wolff, J. Char, X. Frogneux and T. Cantat, Angew.
Chem., Int. Ed., 2017, 56, 5616.
11 Y. Hatanaka and T. Hiyama, Chem. Lett., 1989, 2049.
12 M. Tiwari, Vinit and C. M. Ramachandran, Chem. Phys. Lett., 2019,
726, 111.
13 G. Pelzer, J. Herwig, W. Keim and R. Goddard, Russ. Chem. Bull.,
1998, 47, 904.
Hammett plot (Fig. 2). Hartwig et al. have shown that the
formation of C–S bonds by reductive elimination from Pd(II)
thiolate complexes are, on the contrary, facilitated with electron-
deficient aryl partners.20 The inverse trends may be attributed to
the lower nucleophilicity of the sulfinate anion compared to the
thiolate anion, the lone pair of the sulfur atom being partly
delocalized on the two oxygen atoms (Fig. 2).
In conclusion, we have developed a practical palladium-
catalyzed synthesis of allyl aryl sulfones from readily available
organosilanes, aryl halides, and sulfur dioxide. This process
represents the first approach that introduces a sp2-hybridized
electrophile with an organosilane in a sulfonylative cross-coupling.
14 H. Woolven, C. Gonzalez-Rodriguez, I. Marco, A. L. Thompson and
M. C. Willis, Org. Lett., 2011, 13, 4876.
15 Y. Hatanaka, K.-i. Goda and T. Hiyama, Tetrahedron Lett., 1994,
35, 6511.
16 Y. Hatanaka, S. Fukushima and T. Hiyama, Tetrahedron, 1992, 48, 2113.
Experimental and theoretical investigations have demonstrated 17 C. Amatore, L. Grimaud, G. Le Duc and A. Jutand, Angew. Chem.,
2014, 126, 7102.
the key role of transient sulfinate anions and their coupling
with the aryl halides, mediated by palladium, and they high-
18 A rate-determining reductive elimination is in addition compliant
with the higher efficiency of Xantphos as a ligand (Table 1 and Table
light the challenges facing the synthesis of diaryl sulfones from
organosilanes.
The authors acknowledge for financial support of this
work: CEA, CNRS, CINES (project sis6494), the CHARMMMAT
Laboratory of Excellence, and the European Research Council
(ERC Consolidator Grant 818260 to T.C.).
S1, ESI†). Indeed, because of the steric pressure induced by its wide
bite angle (1111), Xantphos is known to promote difficult reductive
eliminations, notably with weakly nucleophilic partners. See:
(a) P. J. C. Kamer, P. W. N. M. Van Leeuwen and J. N. H. Reek,
Acc. Chem. Res., 2001, 34, 895; (b) V. V. Grushin and W. J. Marshall,
J. Am. Chem. Soc., 2006, 128, 12644; (c) K. Fujita, M. Yamashita,
F. Puschmann, M. M. Alvarez-Falcon, C. D. Incarvito and J. F.
Hartwig, J. Am. Chem. Soc., 2006, 128, 9044.
19 Troupel et al. showed that oxidative addition is facilitated with
electron-poor electrophiles (r 4 0): J.-F. Fauvarque, F. Pflu¨ger and
M. Troupel, J. Org. Chem., 1981, 208, 419.
20 G. Maan, D. Baranano, J. F. Hartwig, A. L. Rheingold and I. A. Guzei,
J. Am. Chem. Soc., 1998, 120, 9205.
Conflicts of interest
There are no conflicts to declare.
This journal is ©The Royal Society of Chemistry 2019
Chem. Commun., 2019, 55, 12924--12927 | 12927